Home

trendEcon Austria

This dashboard is based on the work of trendEcon. The data are scraped from Google Trends through the use of the trendecon R-package. The sectoral indicators are slightly adapted to match Austrian search keywords (e.g. different stores). The main focus of this project is to test the accuracy of the main indicator (PES) in a VAR forecasting framework.

Row

Perceived Economic Situation

Info

Description

The indicator for Perceived Economic Situation includes search terms that reflect people’s worries about the economy. For instance, people then google “economic crisis” (Wirtschaftskrise).

Keywords
  • Wirtschaftskrise
  • Kurzarbeit
  • arbeitslos
  • Insolvenz

Download data

Row

Clothing and Shoes

Info

Description

The category Clothing and Shoes includes clothing and shoe stores as well as a general search terms related to buying clothes and shoes. trendEcon found that people directly google for the brands they like. Note: the searchword “zalando” was not included because Zalando was not available in Austria before 2012.

Keywords
  • mango
  • zara
  • H&M
  • blue tomato
  • schuhe kaufen
  • deichmann

Download data

Row

Home Office

Info

Description

The category Home Office combine Google inquiries related to the new normal of working from home. The working-from-home-index contains the search terms:

Keywords
  • headset
  • monitor
  • maus
  • hdmi

Download data

Row

Gardening and Home Improvement

Info

Description

The category Gardening and Home Improvement includes stores selling materials for home improvement such as building materials, garden accessories and electrical supplies.

Keywords
  • Heim+Hobby
  • Bau+Hobby
  • Bauhaus
  • hornbach
  • obi

Download data

Row

Food Delivery

Info

Description

The category Food Delivery includes search terms related to take away and ordering pizza.

Keywords
  • take away
  • takeaway
  • pizza bestellen

Download data

Row

Cultural

Info

Description

The category Cultural Events includes search terms related to concerts, theatres, cinema and ticket providers for such events.

Keywords
  • kino
  • theater
  • cinema
  • cineplexx
  • oper
  • konzert
  • oeticket

Download data

Row

Travel Abroad

Info

Description

The category Travel Abroad includes search terms used to book flights and holidays.

Keywords
  • städtetrip
  • flug buchen
  • günstige flüge

Download data

Row

Mobility

Info

Description

The category Mobility includes search terms related to ground transportation: For instance, checking the railway schedule or calling a taxi.

Keywords
  • Fahrplan
  • taxi
  • sixt
  • google maps

Download data

Row

Watches and Jewelry

Info

Description

The category Watches and Jewellery includes stores and brands selling luxurious watches and jewellery goods and related, more general search terms for luxury consumer goods.

Keywords
  • juwelier
  • swarovski
  • uhr
  • uhren
  • christ
  • feichtinger

Download data

PES

The main indicator shows co-movement with the consumer confidence and with the quarterly GDP growth rate. However, in contrast to the trendEcon results for Germany and Switzerland, the main indicator as well as the consumer confidence insufficiently capture the impact of the World Financial Crisis on GDP in the years after 2008/09.

VAR

Results

Vector Autoregressive Model with 12 lags: \(\mathbf{x_t^{AT}} = [ PES_t^{AT}, u_t^{AT}, P_t^{AT}]\)

VAR(12)

dependent variables: pes, une, cpi

pes une cpi
(1) (2) (3)
pest-1 0.424*** -0.177*** 0.012
(0.090) (0.050) (0.038)

Δ unet-1

0.026 -0.221*** 0.072
(0.138) (0.077) (0.058)
cpit-1 -0.055 0.089 0.077
(0.175) (0.098) (0.074)
pest-2 0.025 -0.162*** 0.026
(0.099) (0.055) (0.042)

Δ unet-2

-0.165 0.060 0.114*
(0.139) (0.077) (0.059)
cpit-2 0.034 0.157 0.019
(0.178) (0.099) (0.075)
pest-3 0.066 -0.049 0.012
(0.102) (0.057) (0.043)

Δ unet-3

0.032 -0.096 0.016
(0.153) (0.085) (0.065)
cpit-3 0.033 0.066 0.080
(0.174) (0.097) (0.074)
pest-4 0.009 0.209*** 0.026
(0.102) (0.057) (0.043)

Δ unet-4

0.165 -0.156* 0.001
(0.158) (0.088) (0.067)
cpit-4 0.139 -0.008 -0.098
(0.174) (0.097) (0.074)
pest-5 0.033 0.082 0.022
(0.108) (0.060) (0.046)

Δ unet-5

-0.076 -0.202** -0.028
(0.166) (0.092) (0.070)
cpit-5 -0.138 0.077 0.038
(0.179) (0.100) (0.076)
pest-6 0.114 -0.147** -0.019
(0.108) (0.060) (0.046)

Δ unet-6

0.109 0.039 0.009
(0.167) (0.093) (0.071)
cpit-6 -0.194 0.144 0.173**
(0.180) (0.100) (0.076)
pest-7 0.031 0.039 -0.036
(0.110) (0.061) (0.047)

Ε unet-7

0.124 -0.061 -0.133*
(0.166) (0.093) (0.070)
cpit-7 -0.188 -0.015 0.096
(0.196) (0.109) (0.083)
pest-8 0.010 -0.021 -0.068
(0.109) (0.061) (0.046)

Ζ unet-8

0.067 -0.048 -0.114*
(0.161) (0.090) (0.068)
cpit-8 0.001 0.039 -0.161*
(0.197) (0.110) (0.083)
pest-9 0.024 0.013 -0.023
(0.107) (0.060) (0.045)

Η unet-9

0.188 0.119 -0.049
(0.142) (0.079) (0.060)
cpit-9 -0.165 0.024 0.078
(0.200) (0.112) (0.085)
pest-10 0.049 -0.069 -0.112**
(0.106) (0.059) (0.045)

⎖ unet-10

0.062 0.133* -0.100
(0.143) (0.080) (0.061)
cpit-10 -0.192 0.206* -0.060
(0.202) (0.112) (0.085)
pest-11 -0.029 0.157*** 0.122***
(0.105) (0.058) (0.044)

⎖ unet-11

0.020 -0.042 -0.176***
(0.138) (0.077) (0.058)
cpit-11 -0.072 -0.017 0.117
(0.197) (0.110) (0.083)
pest-12 -0.040 0.124** -0.152***
(0.104) (0.058) (0.044)

⎖ unet-12

0.041 -0.452*** -0.136**
(0.129) (0.072) (0.055)
cpit-12 0.112 0.040 0.586***
(0.198) (0.110) (0.084)
constant 0.085 -0.130 0.039
(0.145) (0.081) (0.061)
Observations 163 163 163
R2 0.300 0.563 0.678
Adjusted R2 0.100 0.438 0.586
Residual Std. Error (df = 126) 0.605 0.337 0.257
F Statistic (df = 36; 126) 1.502* 4.502*** 7.372***
Note: p<0.1; p<0.05; p<0.01

Summary

Data

pes - Perceived Economic Situation (Nowcasting proxy for GDP)
Dune - Unemployment rate
cpi - Inflation rate year-on-year

all variables are on a monthly frequency

Model diagnostics - residuals

(p.value threshold = 0.05)

multivariate test for serially correlated errors (Portmanteau- and Breusch-Godfrey): H0 of no serial correlation: accepted. p = 0.33

multivariate Jarque-Bera tests:

H0 of Normality: rejected. p = 0

H0 of Skewness = 0: rejected. p = 0

H0 of Kurtosis (excess curtosis = 0): rejected. p = 0

multivariate ARCH-LM test:

H0 of homoskedasticity: accepted. p = 0.84

Johansen Procedure for VAR:

H0 of no cointegration: rejected. Number of cointegration relationships: 2

Model stability - structural breaks

The data lies between the confidence level boundaries and therefore indicates that there are no structural changes in the data.

Sources

PES - scraped from Google Trends using the trendecon r package.
unemployment - Source: Statistik Austria
consumer price index - Source: Statistik Austria
consumer confidence index - Source: OeNB
gdp - Source: OECD

R-Packages: dygraphs, flexdashboard, htmlwidgets, knitr, tsbox, tidyverse, vars, ggplot2, plotly ggfortify, stargazer, forecast

Forecasting

Row

Out-of sample VECM(12) Forecasting (3M ahead)

Row

In-sample Forecast (3M ahead, 3 rolls) - PES

In-sample Forecast (3M ahead, 3 rolls) - Unemployment

In-sample Forecast (3M ahead, 3 rolls) - CPI

Accuracy

Evaluation of in-sample forecast accuracy (1M ahead)
var ME RMSE MAE MPE MAPE
pes pes 0.559 0.606 0.559 382.011 1022.841
Dune Dune -0.755 0.809 0.755 38.374 66.317
cpi cpi 0.476 0.509 0.476 135.352 135.352
global global 0.093 0.653 0.597 185.246 408.170

consumer confidence

Row

In-sample Forecast (3M ahead, 3 rolls) - consumer confidence

In-sample Forecast (3M ahead, 3 rolls) - Unemplyoment

In-sample Forecast (3M ahead, 3 rolls) - CPI

Accuracy

Evaluation of in-sample forecast accuracy - consumer confidence (3M ahead)

consumer confidence Forecast accuracy:

var ME RMSE MAE MPE MAPE
cc cc 1.001 1.003 1.001 Inf Inf
Dune Dune -0.574 0.664 0.574 40.656 49.975
cpi cpi 0.468 0.584 0.472 125.841 127.183
global global 0.298 0.772 0.682 Inf Inf

crisis period

Row

In-sample Forecast (3M ahead, 3 rolls) - PES

In-sample Forecast (3 ahead, 3 rolls) - Unemplyoment

In-sample Forecast (3M ahead, 3 rolls) - CPI

Accuracy

Evaluation of in-sample forecast accuracy - consumer confidence (3M ahead)

consumer confidence Forecast accuracy:

var ME RMSE MAE MPE MAPE
pes pes -2.245 4.017 2.391 104.818 104.818
Dune Dune 0.263 0.370 0.263 41.245 58.315
cpi cpi -0.093 0.219 0.215 -Inf Inf
global global -0.692 2.332 0.956 -Inf Inf

comparison

Row

Benchmark models (3M ahead, 3 rolls)